The long term goal of this project is to identify the patients with acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS) who are the most likely to respond to decitabine therapy. Decitabine is a hypomethylating agent that has efficacy in both MDS and AML. It can be given as an outpatient, and it is well tolerated in most patients. However, response rates are modest; even with modern aggressive schedules, only 4 5 % of patients achieve a complete response. The molecular basis of decitabine sensitivity and/or resistance is not yet clear. Specific Aims: Aim1. We will define the molecular signature of decitabine responders. We will prospectively bank 125 properly consented patients treated with the current state-of-the-art decitabine protocol. We will comprehensively define patient-specific molecular signatures through exome sequencing and expression profiling, using both mRNA and miRNA based arrays. We will correlate genotyping and expression results with clinical features, including responsiveness to decitabine therapy. These studies will correlate genomic signatures of DNMT3A, I D H I , IDH2, and TET2 with outcomes. In addition, comprehensive, unbiased analysis will determine whether specific molecular signatures are associated with decitabine responses. Aim 2. We will determine whether the rate of AML clearance and persistence of AML-associated subclones corresponds to drug metabolism, molecular, and/or clinical features of AML in each case. We will assess the velocity of patient-specific mutation clearance on day 0, 10, and 28, and the persistence of AML-associated subclones despite blast clearance. We will correlate this with steady-state decitabine drug levels, the reduction of methylcytosine in the total marrow sample (a biomarker of effective dosing), and with clinical response rates and event-free survival.